Artificial intelligence algorithms and in particular machine learning are more and more used at the LMCE to tackle our research problems. In nuclear physics, there are three possible areas of use.

The first one is to accelerate high numerical cost calculations preformed to solve our theoretical models or to emulate (meta-model) of these models directly. Our efforts in this direction focus on the emulation of calculations of binding energy of nuclei and properties associated with their deformation (e.g. their vibrational inertia), as well as on algorithms for the optimization of numerical parameters for mean field calculations. Today, these studies rely heavily on the use of Gaussian processes and neural networks.

Secondly, artificial intelligence is becoming increasingly important in analysing the experimental campaigns conducted at the LMCE. Work is typically underway to to apply classification algorithms to the increasing amount of data recorded during our experiments. An exemple of application is the improvement of the particle discrimination resolution in data collected by our detector devices.

Finally, we are exploring the use of machine learning overcome some pathologies of our current theoretical approaches in the description of nuclear processes. A project in this sense aims at taking advantage of generative algorithms to feed the time-dependent generative coordinate method (TDGCM), a framework to simulate the dynamics of nuclear fission.

All this work has led to new collaborations, notably with ENS Paris-Saclay (Saclay) and the startup Magic LEMP.